Angewandte
Chemie
which coordinate three ZnI2 centers and generate the infinite
network. The formyl group of 3 in the as-synthesized network
1 is located in pore B, and disordered nitrobenzene molecules
fill both pores. As nitrobenzene can react with organozinc
reagents, the included nitrobenzene was replaced with inert
toluene after immersion for 2 d. Complete solvent exchange
was confirmed by X-ray analysis. Pores A and B contained
only toluene molecules, but now the formyl group of 3 was
found in the both pores A and B with 47 and 53%
occupancies, respectively (Figure 1b; see also the Supporting
Information).
The addition of organozinc reagents to aldehyde 3 trapped
within the crystalline flask was examined on large scale.
Crystals of 1 (205 mg) were immersed in a toluene solution of
dimethylzinc (2.0m, 3.0 mL) for 24 h at room temperature.
Excess dimethylzinc toluene solution was removed by decant-
ation and the crystals decomposed with 5.0m hydrochloric
acid. The product was extracted with CH2Cl2 and purified by
column chromatography to give methyl adduct 4a in 85%
yield. Independently, multiple small-scale reactions were
quenched at various time intervals, extracted, and analyzed
by NMR. Plotting the conversion ratio versus time revealed
that the reaction is essentially complete in fewer than 12 h.
Diethylzinc also reacted with aldehyde 3 in network 1 (77 mg)
and gave the ethyl adduct 4b in 90% isolated yield. Micro-
scopic FT-IR spectroscopy of a single crystal of reacted 1
A and several solvent molecules were also present. Despite
the Lewis acidic character and reactivity of alkylzinc reagents,
the framework of 1 was not damaged. The ZnI2 atoms
incorporated into the framework remained intact even
though disproportionation between R2Zn and ZnI2 rapidly
occurs in solution.[8]
We next attempted the one-pot transformation of 3 into
acetate 5a.[9] Crystals containing in situ formed 4a were
transferred to an acetic anhydride/toluene (1:1) solution.
=
After immersion for 24 h at room temperature, a single C O
stretching vibration at 1734 cmÀ1 was observed by micro-
scopic FTIR and assigned to the acylated product 5a. The
crystallinity of the robust network persisted and was con-
firmed by visual observation (Figure 2) and by X-ray dif-
=
showed the disappearance of the strong C O stretching
vibration of the formyl group at 1697 cmÀ1 and corroborated
that the reaction did indeed occur inside the crystals.
Figure 2. The one-pot transformation of aldehyde 3 into acetate 5a
within crystals of 1. a) The reaction. b–d) Microscope images of
crystals of 1 b) before the reaction, c) after treatment with dimethyl-
zinc, and d) after subsequent reaction with acetic anhydride.
Despite the lack of an activating Lewis base, such as
amines or alcohols, the addition of alkylzinc reagents to
aldehydes trapped within the pores led to enhanced reactivity
and selectivity. Outside the crystalline flasks, the reactivity of
3 and dimethylzinc in toluene solution was negligible (5%
yield) and the addition of ZnI2 and/or ligand 2 to the toluene
reaction mixture made little difference (6–8% yield). The
reaction did proceed with diethylzinc (98% conversion) but a
4:6 mixture of 4b and the reduced alcohol (ArCH2OH) was
obtained. The enhanced reactivity and selectivity are pre-
sumably ascribed to high local concentrations and decreased
mobility of the substrate/reagent, respectively, within the
crystal. Furthermore, when diphenylzinc was employed as an
organozinc reagent, aldehyde 3 was quantitatively recovered.
This reagent selectivity can be ascribed to steric repulsion
between bulky zinc reagent and the framework of 1.
X-ray crystallographic analysis of the crystals immediately
following the reaction revealed the robust network frame-
work, but the final product structure was not fully solved
because of strong residual electron densities in the pore;
presumably from unreacted dimethylzinc, the unquenched
zinc alkoxide, and partially hydrolyzed alcohol. The post-
reaction crystals were immersed in toluene (not anhydrous) to
wash away excess zinc reagent and byproduct and to quench
the zinc alkoxide. X-ray analysis of the washed crystals was
successful and the structure of 4a fully solved (Figure 1c).
Washing the crystals with toluene removed the excessive
residual electron density from the pores, and the 1-hydrox-
yethyl group of 4a was clearly observed in pore A. The methyl
adduct is quite large, yet enough space remained within pore
fraction. The crystals containing 5a were decomposed with
acid and extracted, and acetate 5a was confirmed by 1H NMR
(> 95% conversion). In a similar manner, two-step reaction
with diethylzinc and acetic anhydride gave (1-acetoxy)propyl
analogue 5b in 93% yield (see the Supporting Information).
In summary, we demonstrated that even organometallic
reactions proceed smoothly in a single-crystal-to-single-crys-
tal fashion within the pores of porous coordination networks.
These “crystalline molecular flasks” can tolerate highly
reactive organometallic reagents while enhancing both the
reactivity and selectivity. Furthermore, we envision the
applications of crystalline flasks for the design of new
reactions and the detailed structural and mechanistic exami-
nation of organometallic transformations through in situ
crystallography.
Received: April 7, 2010
Revised: May 28, 2010
Published online: July 13, 2010
Keywords: metal–organic frameworks ·organozinc compounds ·
.
porous networks · single-crystal modifications · X-ray diffraction
Angew. Chem. Int. Ed. 2010, 49, 5750 –5752
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
5751